摘要
与传统砂浆相比,新型砂浆具有成分复杂的特点,砂浆性能随成分变化产生的波动性很大。砂浆成分的微小改变导致新拌砂浆性能和硬化砂浆力学性能较大改变,传统砂浆配合比计算方法不适用橡胶粉砂浆。人工神经网络技术通过一组试验数据的学习,使其可以预测橡胶粉砂浆的性能,并以另三组独立的试验数据来检验网络的学习效果,此项研究提供了人工智能在砂浆配合比设计中的应用方法,并为新型砂浆外加剂掺量选择提供另一种手段。
Composition of new-style moriar,compared with the traditional mortar,is more complicated.The properties of new-style mortar fluctuate evidently with the changes of the composition,which result in the traditional mortar can not apply in new-style mortar.The artificial neural network should be trained by a group of data.And then the properties of rubber mortar,compared with those that come from experiments, are predicated.A method of mix proportion design and means for the selection of admixture content are provided and presented.
出处
《混凝土》
CAS
CSCD
北大核心
2008年第5期109-111,共3页
Concrete
基金
福建省自然科学基金计划资助项目(E0610020)
建设部2006年科技计划项目(06-k7-35)
关键词
橡胶粉砂浆
抗压强度
砂浆稠度
BP神经网络
rubber mortar
compressive strength
mortar consistency
BP neural network